Spanning Tree Protocol (STP) allows redundant LAN (Local Area Network) links to be used while preventing frames from looping around the LAN indefinitely through the redundant links. In the implementation of the protocol, loops are discovered before the links comprising them are opened for use, and redundant links are placed in a blocking or standby state to prevent frame forwarding. In the event that a forwarding port becomes disabled or fails in some manner, the Spanning Tree Algorithm is reapplied to recalculate a new topology so that blocked connections can be made operational by activating links that were previously in a blocked or standby state.
In the implementation of a network of DSLAMs (Digital Subscriber Line Access Modules) that are connected to a bridged CO (Central Office) over a ring, STP is used to manage the loop that arises when a ring topology is used. Referring to FIG. 1 there may be seen an Ethernet switch 100 connected to an IP Network 110. Ethernet switch 100 also connects to links 101 and 103 via Ethernet ports 102 and 104 respectively. A plurality of DSLAMs are connected to the Ethernet switch 100. DSLAM 120 connects to link 101 via port 122 and DSLAM 140 connects to link 103 via port 144. DSLAM 120 also connects via port 124 to link 105, and DSLAM 140 connects to link 107 via port 142. Further, DSLAM 130 connects to link 105 via its port 132, and to link 107 via its port 134. Further, each DSLAM connects to a group of digital broadband subscribers, namely DSLAM 120 to subscribers 150a-150d, DSLAM 130 to subscribers 160a-160d, and DSLAM 140 to subscribers 170a-170d. 
It may be seen that links 101, 103, 105 and 107 define a switched loop configuration, connecting DSLAMS 120, 130 and 140 to Ethernet switch 100. The loop configuration provides link redundancy, as should any single link fail, the DSLAM connected to the link may still establish a connection to the Ethernet switch 100 via the port connected to the non-failed link.
When the Spanning Tree Algorithm is implemented upon this switched loop, one of the DSLAM ports will be redundant and placed in a blocked or standby state. Upon a link failure, a re-computing of the spanning tree structure will be initiated, and the unused port will be recruited into the new Spanning Tree.
Variants of SPT include RSPT (Rapid Spanning Tree Protocol) and MSTP (Multiple Spanning Tree Protocol). RSPT provides for faster spanning tree computation after a topology change; whereas STP requires on the tens of seconds (typically 30 to 50 seconds) RSTP is typically able to respond to changes within a second. MSTP is an extension of RSTP to Virtual Local Area Networks (VLANs). This protocol computes a separate spanning tree for each VLAN group and blocks the loops which are redundant within each spanning
Subscriber management is concerned with enforcing subscriber Quality of Service (QoS) and accounting profiles over a service VLAN (in the VLAN per service model) or subscriber VLAN (in the VLAN per subscriber model). Ingress and egress subscriber traffic need to be queued at the proper port leading the subscriber. As well, operational statistics for each subscriber needs to be maintained in order to implement subscriber management policies.
In the event of a port failure, the network will reconfigure via the Spanning Tree Protocol. However, any new port recruited by the protocol will be starting from zero regarding operational statistics, thus leading to potential
One possible remedy for this would be to utilize Link Aggregation Groups (LAGs) in the links. Referring to FIG. 2 [note: like references denote like elements in the following Figures] there may be seen a plurality of DSLAMs 220, 230 and 240 connected to an Ethernet switch 200 in a loop configuration. The number of links comprising the loop configuration has been doubled over the network depicted in FIG. 1, also requiring double the number of ports at the Ethernet switch 200 and the DSLAMs. A Link Aggregation Group may be defined grouping links together, for example link 201a and 201b. This is normally done to provide more bandwidth, but could also serve to enhance subscriber traffic management reliability, as if a single port fails, the remaining port within the LAG can still provide the necessary history and statistics.
A disadvantage of using LAGs to obtaining this increased reliability may be the increased cost in terms of links and ports that need to be provisioned to support LAG operation.